Understand
In this report we describe the technical details of our submission to the EPIC-Kitchens 2019 action recognition challenge.
- To participate in the challenge we have developed a number of CNN-LSTA [3] and HF-TSN [2] variants, and submitted predictions from an ensemble compiled out of these two model families.
- Our submission, visible on the public leaderboard with team name FBK-HUPBA, achieved a top-1 action recognition accuracy of 35.54% on S1 setting, and 20.25% on S2 setting.
Built on
Scaling Egocentric Vision: The EPIC-KITCHENS Dataset
D. Damen, H. Doughty, G. Maria Farinella, S. Fidler, A. Furnari, E. Kazakos, D. Moltisanti, J. Munro, T. Perrett, W. Price, and M. Wray · 2018
Earlier work this paper cites.
Attention is All We Need: Nailing Down Object-centric Attention for Egocentric Activity Recognition
S. Sudhakaran and O. Lanz · 2018
Earlier work this paper cites.
Similar
Hierarchical Feature Aggregation Networks for Video Action Recognition
S. Sudhakaran, S. Escalera, and O. Lanz · 2019
Cited alongside, same era.
Then
LSTA: Long Short-Term Attention for Egocentric Action Recognition
S. Sudhakaran, S. Escalera, and O. Lanz · 2019
Closest in time.
Beyond the bibliography
alphaXiv searches the wider corpus for related work and actual follow-ups.
alphaXiv is searching for related work…